Experimental realization of a dynamic squeezing gate

Abstract

Squeezing is a nonlinear Gaussian operation that is the key component in construction of other nonlinear Gaussian gates. In our implementation of the squeezing gate, the amount and the orientation of the squeezing can be controlled by an external driving signal with 1 MHz operational bandwidth. This opens a brand new area of dynamic Gaussian processing. In particular, the gate can be immediately employed as the feed-forward needed for the deterministic implementation of the quantum cubic gate, which is a key piece of universal quantum information processing.PACS numbers: 03.67. Lx, 42.50.Dv, 42.50.Ex, Quantum information processing with continuous variable systems (CV) has many tools. They could be divided into two broad categories -Gaussian and nonGaussian. The Gaussian tools comprise of Gaussian quantum states that can be represented by a Gaussian Wigner function, of Gaussian measurements that project on Gaussian states, and of Gaussian operations that transform Gaussian states into different Gaussian states [1]. The non-Gaussian tools category then includes everything else. The non-Gaussian category is much broader and much more powerful. There are many quantum information protocols that cannot be implemented with Gaussian tools alone, quantum computation [2, 3], entanglement distillation [4][5][6], and error correction [7] are just the three most prominent examples.As a consequence, there is an understandable thirst for all matters non-Gaussian. In quantum optics, which is the experimental platform of choice when it comes to tests of CV paradigms [8], the non-Gaussian features need to come from interactions with discrete variable physical systems [9][10][11][12][13][14][15][16][17], or from discrete measurements [1,7,18,19]. These two general approaches also differ with respect to quantum systems for which they can be applied. While the interaction with discrete variable systems is best realized by a standing wave mode in a resonator, the discrete projective measurements work better with traveling light. And here comes another distinction. The traveling modes of light are much more suitable for implementation of Gaussian operations. This is significant, because the non-Gaussian resources are useful only when the Gaussian tools are refined enough to operate without a hitch. To present a specific example, consider the issue of universal quantum information processing. In CV world this means the ability to implement a unitary operation with an arbitrary Hamiltonian [2,20]. For this we need to have access to the cubic operation -a quantum operation with Hamiltonian composed of third power of quadrature operators -as well as the complete range of Gaussian operations.The Gaussian states, operations, and measurements are the foundations on which the CV quantum information processing is built. Homodyne detection, squeezed states and Gaussian linear operations in the form of displacement and passive linear optics are already staples of the contemporary experimental practice. The measurement induced paradigm [21], whi...

Implementation of a quantum cubic gate by an adaptive non-Gaussian measurement

“…Furthermore, they have been tested on non-Gaussian states of light [18], to prove their general applicability. Recently, the quadratic electrooptical feedforward control has been demonstrated [19]. In addition, to independently obtain the cubic state, a finite dimensional approximation of the cubic state has been suggested [20] and its performance in the GKP scheme has been discussed.…”

“…The topology will be then similar to the simple one used for a measurement-induced squeezing gate [14,[17][18][19]25]. Non-Gaussian operations can be realized by simply substituting a measurement of nonlinear combination of quadrature amplitudes for the Gaussian homodyne measurement [26,27].…”

“…While each of them has been already demonstrated [15,19], the total implementation to build a unitary cubic operation demands three squeezed states as well as one non-Gaussian ancilla, and is probably not the simplest arrangement. In contrast, we here use adaptability of linear optical schemes and propose a better and simpler topology with linear optics and suitable ancillary states.…”

Quantum simulation of quantum field theory using continuous variables

“…From its original conception in 1999 [5], progress began to accelerate after a cluster state [6] version was established in 2006 [7,8], leading to something significantly more tangible for experimentalists. This resulted in numerous proof-of-principle demonstrations [9][10][11][12], currently culminating in a 10,000 node cluster [13] created 'onthe-go' along with a 60 node cluster created simultaneously [14]. From a theoretical perspective, much progress has been made [15][16][17][18][19][20][21][22][23], including recently, an important fault tolerant architecture [24], achieved by leveraging the Gottesman-Kitaev-Preskill (GKP) encoding [25].…”

Quantum Machine Learning over Infinite Dimensions

“…Seven years later, the cluster state version [28] of CVs [29,30], accelerated the field due to experimental interest. The result were proof-of-principle demonstrations [31][32][33][34], which culminated in a time domain one-million-node cluster [35,36] and a 60-node frequency domain cluster [37]. Further important theoretical work was also carried out [38][39][40][41][42][43][44][45][46][47], including an important CV architecture that was fault tolerant [48].…”